• ETRI Journal
• /
• 제45권5호
• /
• pp.836-846
• /
• 2023

Mobile robots are used in modern life; however, object recognition is still insufficient to realize robot navigation in crowded environments. Mobile robots must rapidly and accurately recognize the movements and shapes of pedestrians to navigate safely in pedestrian-rich spaces. This study proposes real-time, accurate, three-dimensional (3D) multi-pedestrian detection and tracking using a 3D light detection and ranging (LiDAR) point cloud in crowded environments. The pedestrian detection quickly segments a sparse 3D point cloud into individual pedestrians using a lightweight convolutional autoencoder and connected-component algorithm. The multi-pedestrian tracking identifies the same pedestrians considering motion and appearance cues in continuing frames. In addition, it estimates pedestrians' dynamic movements with various patterns by adaptively mixing heterogeneous motion models. We evaluate the computational speed and accuracy of each module using the KITTI dataset. We demonstrate that our integrated system, which rapidly and accurately recognizes pedestrian movement and appearance using a sparse 3D LiDAR, is applicable for robot navigation in crowded spaces.

• Journal of the Optical Society of Korea
• /
• 제18권5호
• /
• pp.605-610
• /
• 2014

Clouds play an important role in climate change, in the prediction of local weather, and also in aviation safety when instrument assisted flying is unavailable. Presently, various ground-based instruments used for the measurements of the cloud base height or velocity. Lidar techniques are powerful and have many applications in climate studies, including the clouds' temperature measurement, the aerosol particle properties, etc. Otherwise, it is very circumscribed in cloud velocity measurements because there is no Doppler effect if the clouds move in the perpendicular direction to the laser beam path of Doppler lidar. In this paper, we present a method for the measurement of cloud velocity using lidar's range detection and DIC (Digital Image Correlation) system to overcome the disadvantage of Doppler lidar. The lidar system acquires the distance to the cloud, and the cloud images are tracked using the developed fast correlation algorithm of DIC. We acquired the velocities of clouds using the calculated distance and DIC algorithm. The measurement values had a linear distribution.

• Journal of Information Processing Systems
• /
• 제20권3호
• /
• pp.375-390
• /
• 2024

Edge computing architecture has effectively alleviated the computing pressure on cloud platforms, reduced network bandwidth consumption, and improved the quality of service for user experience; however, it has also introduced new security issues. Existing anomaly detection methods in big data scenarios with cloud-edge computing collaboration face several challenges, such as sample imbalance, difficulty in dealing with complex network traffic attacks, and difficulty in effectively training large-scale data or overly complex deep-learning network models. A lightweight deep-learning model was proposed to address these challenges. First, normalization on the user side was used to preprocess the traffic data. On the edge side, a trained Wasserstein generative adversarial network (WGAN) was used to supplement the data samples, which effectively alleviates the imbalance issue of a few types of samples while occupying a small amount of edge-computing resources. Finally, a trained lightweight deep learning network model is deployed on the edge side, and the preprocessed and expanded local data are used to fine-tune the trained model. This ensures that the data of each edge node are more consistent with the local characteristics, effectively improving the system's detection ability. In the designed lightweight deep learning network model, two sets of convolutional pooling layers of convolutional neural networks (CNN) were used to extract spatial features. The bidirectional long short-term memory network (BiLSTM) was used to collect time sequence features, and the weight of traffic features was adjusted through the attention mechanism, improving the model's ability to identify abnormal traffic features. The proposed model was experimentally demonstrated using the NSL-KDD, UNSW-NB15, and CIC-ISD2018 datasets. The accuracies of the proposed model on the three datasets were as high as 0.974, 0.925, and 0.953, respectively, showing superior accuracy to other comparative models. The proposed lightweight deep learning network model has good application prospects for anomaly traffic detection in cloud-edge collaborative computing architectures.

• 정보과학회 논문지
• /
• 제44권10호
• /
• pp.1112-1123
• /
• 2017

최근 다수의 콘텐츠 서비스 제공자가 제공하는 콘텐츠 중심 서비스가 클라우드로 이전함과 동시에 온라인 상의 유사 중복 콘텐츠가 급격히 증가함에 따라, 불필요한 과잉 검색 결과를 초래하는 등 클라우드 기반 데이터 검색 서비스의 품질이 저하하고 있다. 또한 데이터 보호법 등에 의거, 각 서비스 제공자는 서로 다른 비밀키를 이용하여 콘텐츠를 암호화하기 때문에 데이터 검색이 어렵다. 따라서, 검색 프라이버시를 보장하면서 유사 중복 데이터 검색의 정확도까지 보장하는 서비스의 구현은 기술적으로 어려운 실정이다. 본 연구에서는, 클라우드 환경에서 데이터 복호 없이 불필요한 검색 결과를 제거함으로써 검색서비스 품질을 제고하며, 동시에 효율성까지 개선된 유사 중복 검출 기법을 제안한다. 제안 기법은 검색 프라이버시와 콘텐츠 기밀성을 보장한다. 또한, 사용자 측면의 연산 비용 및 통신 절감을 제공하며, 빠른 검색 평가기능을 제공함으로써 유사 중복 검출 결과의 신뢰성을 보장한다. 실제 데이터를 통한 실험을 통해, 제안 기법은 기존 연구 대비 약 70.6%로 성능이 개선됨을 보인다.

• 전기전자학회논문지
• /
• 제19권1호
• /
• pp.58-63
• /
• 2015

본 논문에서는 클라우드 기반의 모바일 지능형 관제시스템 개발을 제안하였다. 모바일 지능형 관제시스템은 클라우드 서버, 미들웨어, 센서로 구성되어 있다. 각 모듈들은 모바일 환경에서 제어되고 주변 환경에 대한 각 기기의 동작 상태를 모니터링 할 수 있다. 본 논문에서는 침입자를 감지하기 위해 영상 기반 움직임 감지 알고리즘을 적용하였고, 움직임 감지 실험에서 움직임 검출율이 평균 12.3% 높게 측정되어 보안장치로써의 타당성을 확인하였다.

• International Journal of Computer Science & Network Security
• /
• 제23권10호
• /
• pp.89-96
• /
• 2023

Intrusion detection has been widely studied in both industry and academia, but cybersecurity analysts always want more accuracy and global threat analysis to secure their systems in cyberspace. Big data represent the great challenge of intrusion detection systems, making it hard to monitor and analyze this large volume of data using traditional techniques. Recently, deep learning has been emerged as a new approach which enables the use of Big Data with a low training time and high accuracy rate. In this paper, we propose an approach of an IDS based on cloud computing and the integration of big data and deep learning techniques to detect different attacks as early as possible. To demonstrate the efficacy of this system, we implement the proposed system within Microsoft Azure Cloud, as it provides both processing power and storage capabilities, using a convolutional neural network (CNN-IDS) with the distributed computing environment Apache Spark, integrated with Keras Deep Learning Library. We study the performance of the model in two categories of classification (binary and multiclass) using CSE-CIC-IDS2018 dataset. Our system showed a great performance due to the integration of deep learning technique and Apache Spark engine.

• 해양환경안전학회:학술대회논문집
• /
• 해양환경안전학회 2017년도 공동학술발표회
• /
• pp.229-229
• /
• 2017

• 융합보안논문지
• /
• 제24권2호
• /
• pp.19-29
• /
• 2024

크립토재킹 공격은 암호 화폐 채굴에 필요한 컴퓨팅 자원을 탈취하여 사용자의 가용성을 침해하는 공격이다. 공격의 대상은 일반적인 데스크톱이나 서버 환경에서부터 클라우드 환경까지 점차 다변화되고 있다. 따라서 다양한 컴퓨팅 환경에 적합한 크립토 마이너 탐지 기법의 적용이 필수적이다. 하지만 기존의 탐지 방법론들은 특정 환경에서만 탐지가 시행되었기 때문에 환경별로 적용할 수 있는 방법론에 대해서 비교분석이 제대로 수행되지 않았다. 따라서 본 연구에서는 종래의 크립토 마이너 탐지 기법들에 대한 분류 기준을 수립하고, 각자 다른 실험 환경과 데이터 셋을 기반으로 한 기존의 크립토 마이너 탐지 기법에 대한 심층적인 비교분석을 통해 클라우드 환경에서 적용 가능한 복합적이고 통합적인 탐지 프레임워크를 제시한다.

• 대기
• /
• 제23권4호
• /
• pp.471-483
• /
• 2013

Nighttime sea fog detection from satellite is very hard due to limitation in using visible channels. Currently, most widely used method for the detection is the Dual Channel Difference (DCD) method based on Brightness Temperature Difference between 3.7 and 11 ${\mu}m$ channel (BTD). However, this method have difficulty in distinguishing between fog and low cloud, and sometimes misjudges middle/high cloud as well as clear scene as fog. Using CALIPSO Lidar Profile measurements, we have analyzed the intrinsic problems in detecting nighttime sea fog from various satellite remote sensing algorithms and suggested the direction for the improvement of the algorithm. From the comparison with CALIPSO measurements for May-July in 2011, the DCD method excessively overestimates foggy pixels (2542 pixels). Among them, only 524 pixel are real foggy pixels, but 331 pixels and 1687 pixels are clear and other type of clouds, respectively. The 514 of real foggy pixels accounts for 70% of 749 foggy pixels identified by CALIPSO. Our proposed new algorithm detects foggy pixels by comparing the difference between cloud top temperature and underneath sea surface temperature from assimilated data along with the DCD method. We have used two types of cloud top temperature, which obtained from 11 ${\mu}m$ brightness temperature (B_S1) and operational COMS algorithm (B_S2). The detected foggy 1794 pixels from B_S1 and 1490 pixel from B_S2 are significantly reduced the overestimation detected by the DCD method. However, 477 and 446 pixels have been found to be real foggy pixels, 329 and 264 pixels be clear, and 989 and 780 pixels be other type of clouds, detected by B_S1 and B_S2 respectively. The analysis of the operational COMS fog detection algorithm reveals that the cloud screening process was strictly enforced, which resulted in underestimation of foggy pixel. The 538 of total detected foggy pixels obtain only 187 of real foggy pixels, but 61 of clear pixels and 290 of other type clouds. Our analysis suggests that there is no winner for nighttime sea fog detection algorithms, but loser because real foggy pixels are less than 30% among the foggy pixels declared by all algorithms. This overwhelming evidence reveals that current nighttime sea fog algorithms have provided a lot of misjudged information, which are mostly originated from difficulty in distinguishing between clear and cloudy scene as well as fog and other type clouds. Therefore, in-depth researches are urgently required to reduce the enormous error in nighttime sea fog detection from satellite.

• 대한원격탐사학회지
• /
• 제38권6_1호
• /
• pp.1181-1189
• /
• 2022

구름은 대기 중에 떠 있는 작은 물방울이나 얼음 알갱이들 또는 혼합물 등으로 구성되며 지구 표면의 약 2/3를 덮고 있다. 위성영상내에서의 구름은 일부 다른 지상 물체 또는 지표면과 유사한 반사도 특성으로 인해 구름과 구름이 아닌 영역을 분리하는 구름탐지는 매우 어려운 작업이다. 특히 뚜렷한 특징을 가지는 두꺼운 구름과 달리 얇은 반투명 구름은 위성영상내에서 구름과 배경의 대비가 약하고 지표면과 혼합되어져 나타나기 때문에 대부분 구름탐지에서 쉽게 놓쳐지고 많은 어려움을 주는 대상으로 작용한다. 이러한 구름탐지의 반투명 구름의 한계점을 극복하기 위해, 본 연구에서는 머신러닝 기법(Random Forest [RF], Convolutional Neural Networks [CNN])을 활용하여 반투명 구름을 중점으로 한 구름탐지 연구를 수행하였다. Reference자료로는 MOderate Resolution Imaging Spectroradiometer (MODIS)에서 제공하는 MOD35자료에서 Cloud Mask와 Cirrus Mask를 활용하였으며 반투명 구름 픽셀을 고려한 모델 훈련을 위해 훈련 데이터의 픽셀 비율을 구름, 반투명 구름, 청천이 약 1:1:1이 되도록 구성하였다. 연구의 정성적 비교 결과, RF와 CNN 모두 반투명 구름을 포함한 다양한 형태의 구름 등을 잘 탐지하였고, RF 모델 결과와 CNN 모델 결과를 혼합한 RF+CNN경우에는 개별 모델의 한계점을 개선시키며 구름탐지가 잘 수행되어진 것을 확인하였다. 연구의 정량적 결과 RF의 전체 정확도(OA) 값은 92%, CNN은 94.11%를 보였고, RF+CNN은 94.29%의 정확도를 보였다.